Lichao Sun
Papers
4
Total Citations
10
H-Index
2
About
Lichao Sun is a dynamic researcher at the intersection of reinforcement learning, generative AI, and multi-agent systems, with work spanning both foundational machine learning theory and high-impact applications. His research addresses some of the most pressing challenges in modern AI, including generalization in visual-based reinforcement learning, continual learning, and uncertainty estimation in multi-agent environments. Sun's contributions are notably broad in scope. His work on saliency-guided feature decorrelation tackles the critical problem of agent generalization across unseen environmental variations — a bottleneck in deploying RL systems in real-world settings. His development of the Continual Diffuser (CoD) framework advances offline reinforcement learning by enabling models to adapt across evolving tasks without catastrophic forgetting, a significant step toward lifelong learning systems. In biomedical AI, his Bora model pioneers generalist video generation for medical applications, opening doors for surgical training and data augmentation. His distributional reward estimation research further strengthens multi-agent RL by addressing reward uncertainty in complex cooperative settings. With growing citation impact across multiple domains, Sun represents an emerging voice pushing the boundaries of robust, adaptable, and applied artificial intelligence research.
Research Focus
Key Achievements
Top Papers
- 1Learning Generalizable Agents via Saliency-Guided Features Decorrelation3 citations · 2023
- 2Bora: Biomedical Generalist Video Generation Model3 citations · 2024
- 3
- 4